Search

In this paper, we propose a variance reduction approach for Markov chains based on additive control variates and the minimization of an appropriate estimate for the asymptotic variance. We focus on the particular case when control variates are represented as deep neural networks. We derive the optimal convergence rate of the asymptotic variance under various ergodicity assumptions on the underlying Markov chain. The proposed approach relies upon recent results on the stochastic errors of variance reduction algorithms and function approximation theory. [ABSTRACT FROM AUTHOR]

This paper proposes a novel L1-based predictor–corrector method for the fractional differential equations involving generalized-Caputo type derivative. A decomposition scheme is used to obtain the three-point predictor–corrector formula. The error and stability of the proposed method are given in detail. A computer virus and a five-dimensional Hopfield neural network models are solved using the proposed approach. [ABSTRACT FROM AUTHOR]

In this paper, stochastic predictive computing networks are exploited to investigate the dynamics of the SIS with vaccination impact based epidemic model (SISV-EM) represented by nonlinear systems of stochastic differential equations (SDEs) by exploitation of artificial neural networks (ANNs) with the backpropagated Levenberg-Marquardt technique (BLMT) i.e., (ANNs-BLMT) to approximate the solution behavior. The stochastic nonlinear SISV-EM is governed with three classes: susceptible, infectious, and vaccinated populations. The referenced or target datasets for ANNs-BLMT are constructed by employing Euler-Maruyama (EM) scheme for solving stochastic differential systems in case of sufficiently various nonlinear SISV-EM scenarios by varying the percentage of vaccination for newly born, the coefficient of transmission, the natural mortality rates, the infectious rates of recovery, the rate at which vaccinated people lose their immunity, the rate of death caused by disease, the proportion of vaccinated against susceptible and the white noise in the environment. Based on arbitrary training, testing, and validation samples from the referenced dataset, the ANNs-BLMT provides an approximate solution for the stochastic nonlinear SISV-EM, with significant correlations to the referenced results. Exhaustive simulation-based results using error histograms, mean square errors, and regression analyses further demonstrate that the proposed ANNs-BLMT is efficient, consistent, and accurate for solving SISV-EM. • A two-layer framework of ANNs-BLMT is proposed as an innovative technique based on a stochastic computing paradigm to investigate the dynamics of stochastic nonlinear SISV-EM. • Innovative technique based on a stochastic computing ANNs-BLMT to investigate the dynamics of stochastic nonlinear SISV-EM Reference dataset for ANNs -BLMT is developed with Euler-Maruyama (EM) scheme nonlinear SISV-EM with varying parametersMean square error criteria is used for training of ANNs-BLMT to find approximate solutions to a variety of SISV-EM scenarios Computation of error histogram illustration, regression metrics, and MSE learning curves prove the performance of ANNs-BLMT. • The mean square error criteria are effectively utilized in approximation theory to develop an objective function for training the ANNs-BLMT to determine approximate solutions to a variety of stochastic nonlinear SISV-EM scenarios. • The computation of error histogram illustration, regression metrics, and MSE learning curves significantly improve the performance, precision, and consistency of the ANNs-BLMT for solving the stochastic nonlinear SISV-EM. [ABSTRACT FROM AUTHOR]

This article deals with global μ -stabilization of quaternion-valued inertial bidirectional associative memory neural networks (QVIBAMNNs) with time-varying delays via time-delayed impulsive control. Firstly, the existence and uniqueness of the equilibrium point are proved by the homeomorphism mapping method. Then, based on an improved differential inequality, sufficient conditions for global μ -stabilization of QVIBAMNNs via time-delayed impulsive control are obtained. Significantly, the activation function in this paper is an extension of Lipschitz condition, which is superior to general quaternion-valued activation function. And quaternion is not required to be decomposed into real or complex values in this paper. Ultimately, some numerical simulations have proved the feasibility of the theorems. [ABSTRACT FROM AUTHOR]

Measuring risk is at the center of modern financial risk management. As the world economy is becoming more complex and standard modelling assumptions are violated, the advanced artificial intelligence solutions may provide the right tools to analyse the global market. In this paper, we provide a novel approach for measuring market risk called Encoded Value-at-Risk (Encoded VaR), which is based on a type of artificial neural network, called Variational Auto-encoders (VAEs). Encoded VaR is a generative model which can be used to reproduce market scenarios from a range of historical cross-sectional stock returns, while increasing the signal-to-noise ratio present in the financial data, and learning the dependency structure of the market without any assumptions about the joint distribution of stock returns. We compare Encoded VaR out-of-sample results with twelve other methods and show that it is competitive to many other well-known VaR algorithms presented in the literature. [ABSTRACT FROM AUTHOR]

Machine learning becomes popular and is used for a wide range of problems in various areas of applied sciences. In dynamical systems, machine learning methods are applied to solve differential equations. In this paper, we develop an artificial network to solve systems of ordinary differential equations. For the network, we use a multilayer perceptron networks, which is a fully connected feedforward network to predict the flow of a specific system. In order to improve the long time estimation of the method, we introduced an upper bound control strategy. To deal with stiff ODE systems(such as slow-fast coupled system), a new algorithm, Finite Neural Element method, is introduced. By numerical simulation, the novel algorithm is proved to have better efficiency and accuracy than direct machine learning method. • A deep neural network is developed to solve a bundle of solutions of the differential equations. • A control strategy for the neural network is introduced to increase the accuracy of the method. • The Finite Neural Element method is capable of solving systems with multiple time scales. [ABSTRACT FROM AUTHOR]

This paper proposes a customized hybrid model of artificial neural network (ANN) and genetic algorithms for an efficient diabetes disease prediction framework. Our customized hybrid model uses an improvised technique of detecting the more visible patterns of relations between the variables. Initially, the input medical dataset is preprocessed using a novel normalization technique that works consistently for all degrees of skewness of data. Then, our proposed decision-making algorithm will correctly identify the degree of importance of each variable in influencing the output, and thus priority will be given to the variables that are deemed most important. This is then followed by the implementation of a regularization method that is custom-made for the prediction of diabetes. Such a customized regularization method is considered asymmetrical because the positive numbers are more favored compared to negative numbers, and this was decided based on the characteristics of the dataset. The proposed technique deals with missing numbers as a separate kind of entity compared to numerical entries and can adapt itself to a given dataset. The proposed customized hybrid model and its accompanying decision-making algorithm were applied to the Pima Indian Diabetes dataset sourced from the UCI Machine Learning Repository with an 80% prediction accuracy. [ABSTRACT FROM AUTHOR]

Dynamic metasurface is an emerging concept for achieving a flexible control of electromagnetic waves. Generalized sheet transition conditions (GSTCs) can be used to model the relationship between the electromagnetic response and surface susceptibility parameters characterizing a metasurface. However, when it comes to the inverse problem of designing and controlling a metasurface in a space–time varying context based on GSTCs, the dynamic synthesis of the susceptibility parameters is a difficult and non-intuitive task. In this paper, we transform the inverse problem of solving dynamic susceptibility parameters into a sequence of control problems. Based on FDTD numerical simulations, a Deep Reinforcement Learning (DRL) framework using a proximal policy optimization (PPO) algorithm and a fully connected neural network is designed to control the susceptibility parameters intelligently and efficiently, promoting the further expansion of the application range of metasurface and thus helping realize more flexible and effective control of electromagnetic waves. We provide numerical results in a one-dimensional setting to show the applicability, correctness and effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

The main purpose of the present work is to propose an effective tool which allows to ensure the protection and the safety measures against the instability phenomena in a gas turbine based on the modelling of its dynamic behaviour. In order to provide an efficient diagnostic strategy for this type of rotating machine, a supervision system based on the development of artificial neural network tools is proposed in this paper. Where, the dynamic nonlinear autoregressive approach with external exogenous input NARX is used for the identification of the studied system dynamics, to monitor the vibrational dynamics of the operating turbine. This leads to establishing a solution for the different ranges of rotational speed and ensuring dynamic stability through the vibration indicators, determined by the proposed neural network approach. Also, offer a normalized mean square error on the order of 3.8414e−3 for the high-pressure turbine, 1.29152e−1 for the gas control valve and 2.12090 e-4 for the air control valve. Furthermore, it permits the vibration monitoring and efficiently extracts the essentials of dynamic model behaviour, to effectively size the operating gas turbine system. The obtained results of the application of the proposed approach on the gas turbine system presented in this paper proves its ability for the detection and the management on real-time of the eventual failures caused mainly by intrinsic vibrations. On the other side, these results prove clearly the effectiveness of the use of the artificial neural networks as a very powerful calculation tools in the modelling of complex dynamic systems. [ABSTRACT FROM AUTHOR]

This paper introduces a 3-layer feed-forward neuronet model, trained by novel beetle antennae search weights-and-structure-determination (BASWASD) algorithm. On the one hand, the beetle antennae search (BAS) algorithm is a memetic meta-heuristic optimization algorithm capable of solving combinatorial optimization problems. On the other hand, neuronets trained by a weights-and-structure-determination (WASD) algorithm are known to resolve the shortcomings of traditional back-propagation neuronets, including slow speed of training and local minimum. Combining the BAS and WASD algorithms, a novel BASWASD algorithm is created for training neuronets, and a multi-input BASWASD neuronet (MI-BASWASDN) model is introduced. Using a power sigmoid activation function and while managing the model fitting and validation, the BASWASD algorithm finds the optimal weights and structure of the MI-BASWASDN. Four financial datasets, taken from the European Central Bank publications, validate and demonstrate the MI-BASWASDN model's outstanding learning and predicting performance. Also included is a comparison of the MI-BASWASDN model to three other well-performing neural network models, as well as a MATLAB kit that is publicly available on GitHub to promote and support this research. [ABSTRACT FROM AUTHOR]

This paper deals with self-triggering adaptive optimal control for nonlinear continuous-time systems. We propose a novel self-triggering control structure concerning a special encoding mechanism, which combines the trigger time of control and sampling and reduces both the control time and the sampling time. Such a triggering structure ensures the existence of a maximum triggering time in self-triggering control. When the system expression is known, the encoding mechanism will lead to high quantitative accuracy at a limited channel transmission rate. Moreover, we also provide a new control algorithm and triggering conditions of the proposed structure. Specifically, this algorithm solves the optimal control strategy by using the cost function approximated by neural networks. Besides, the derived closed-loop system is proven to be asymptotically stable. Finally, two examples are provided to illustrate the effectiveness of the proposed control method. [ABSTRACT FROM AUTHOR]

The success of the shooting of a moving vehicle can be achieved by providing barrel stabilization. In the meantime, it is necessary barrel angle must be positioned is instantaneously calculated accurately. In this paper, the required barrel elevation angle was calculated via artificial neural networks and the shooting success to a fixed target from a moving tank was investigated. For this purpose, a mathematical model of a tank, which has four degrees of freedom, with a barrel mounted on it has been utilized. The barrel is moved using an inverted slider mechanism. The system is driven hydraulically by considering the movement of the piston–cylinder pair in the inverted slider mechanism. The conventional PID control method has been preferred for angular position control of the barrel. Simulations have been performed with the MATLAB package program. System responses have been obtained graphically and the results have been presented in tabular form. It is seen that the pressure, flow, etc. obtained from the results overlap with the physical system. It is observed that the elevation angle of 10 degrees is captured and shot very well when the vehicle is stationary. In the moving state of the vehicle, the 10-degree elevation angle capture as well. In the moving case, there are tiny deviations in the target hit but these values can be considered to be reasonable levels. The deviation amount for the target 4 kilometers away is 16.28 meters. [ABSTRACT FROM AUTHOR]

In this paper, a neuro-evolution based numerical computing approach is presented for the solution of nonlinear third order multi-singular Emden–Fowler system of differential equations (MS-EF-SDEs) by manipulating the proficiency of continuous mapping through exploitation of feed-forward artificial neural networks (ANN). The weights or decision variables of these networks are optimized with genetic algorithms (GAs) and sequential quadratic programming (SQP), i.e., ANN-GA-SQP. An error based figure of merit is introduced using the differential model of MS EF-SDE along with corresponding boundary conditions. The objective/cost function is optimized by integrating capability of global and local search with GA and SQP, respectively. The competency of the designed ANN-GA-SQP approach in terms of significance, efficiency and consistency is perceived by solving MS-EF-SDEs. Moreover, statistical based investigations are implemented to validate the correctness of ANN-GA-SQP. [ABSTRACT FROM AUTHOR]

Forecasting of energy-related variables is crucial for accurate planning and management of electrical power grids, aiming at improving overall efficiency and performance. In this paper, an artificial neural network (ANN)-based model is investigated for short-term forecasting of the hourly wind speed, solar radiation, and electrical power demand. Specifically, the non-linear autoregressive network with exogenous inputs (NARX) ANN is considered, compared to other models, and then selected to perform multi-step-ahead forecasting. Different time horizons have been considered in the range between 8 and 24 h ahead. The simulation analysis has put in evidence the main advantage of the proposed method, i.e., its capability to reconcile good forecasting performance in the short-term time horizon with a very simple network structure, which is potentially implementable on a low-cost processing platform. [ABSTRACT FROM AUTHOR]

Abstract In this paper, the multistability and multiperiodicity problems are investigated for the memristor-based fuzzy cellular neural networks (MFCNNs) with nonmonotonic activation functions and unbounded time-varying delays. Based on the fixed point theorem and the geometrical properties of activation functions, sufficient criteria are obtained to ensure such n -neuron MFCNNs can have at least ∏ i = 1 n (2 K i + 1) equilibrium points with K i > 0 in which ∏ i = 1 n (K i + 1) are locally μ -stable. As an extension of the theory, the existence of ∏ i = 1 n (K i + 1) locally exponentially stable periodic solutions with time-periodic inputs is also derived. Finally, one example is presented to confirm our results. Highlights • This paper combines the memristor and fuzzy logic into the neural networks. • The activation functions are nonmonotonic and nonlinear in any open interval. • Both multiple μ -stability and multiperiodicity of neural networks are studied. [ABSTRACT FROM AUTHOR]

In this paper, Artificial Neural Network (ANN) power management for a reverse osmosis desalination unit fed by hybrid renewable energy sources solar PV and wind turbine associated to battery bank as storage element is studied. The ANN power management system has as main objective to ensure the smooth transfer of the generated power by these sources under the variability and intermittency of the wind speed and the irradiation during 24 h of operation considering the limitation constraints of the RO unit and the need water profile. The design, the modeling and the control strategies of all the components are made in this study using Matlab/Simulink. The results show the ability of the ANN power manager to define the operating modes based on the proposed flow chart. [ABSTRACT FROM AUTHOR]

In this paper an Artificial Neural Network (ANN) algorithm is presented in order to solve the problems associated with the conventional DTC approach. In order to improve the performances of the DTC controlled PMSM and to reject the disturbances, an ANN algorithm is used. This intelligent artificial technique is used to select the optimal voltage vector. In order to reduce the torque and flux ripples, the hysteresis comparators and the switching table have been substituted by the ANN technique. Simulation using Matlab/Simulink environment and experimental results around the Dspace-1104, are presented to test the performances of this approach. Simulation and experimental results show the high performances of the ANN-DTC compared to the conventional DTC; in particular the reduction of the ripples in torque and flux. [ABSTRACT FROM AUTHOR]

Abstract This paper focuses on the exponential convergence of impulsive inertial complex-valued neural networks with time-varying delays. The system can be expressed as a first order differential equation by selecting a proper variable substitution. By constructing proper Lyapunov–Krasovskii functionals and using inequality techniques, some delay-dependent sufficient conditions in linear matrix inequality form are proposed to ascertain the global exponential convergence of the addressed neural networks with two classes of complex-valued activation functions. The framework of the exponential convergence ball domain in which all trajectories converge is also given. Meanwhile, the obtained results here do not meet that the derivatives of the time-varying delays are less than one and there are also no limit to the strength of impulses. The methods here can also be applied to deal with multistable and monostable neural networks because of making no hypotheses on the amount of the equilibrium points. Finally, two examples are given to demonstrate the validity of the theoretical results. [ABSTRACT FROM AUTHOR]

Abstract This paper studies the stability problem of impulsive complex-valued neural networks with time delay. Firstly, impulses are viewed as disturbance. Criteria on uniqueness and global exponential stability of equilibrium point are established using the Lyapunov function method. A modified version of Halanay-type inequality is used to make the results in Song et al. (2016) much better. Secondly, impulses are used as control input. Sufficient conditions are derived to ensure the exponential stability of the equilibrium point by using the Razumikhin technique. This result explains the dynamics of example 2 in Song et al. (2016), which does not give any theoretical result on impulsive control problem. Two numerical examples are presented to demonstrate the effectiveness of the theorems. [ABSTRACT FROM AUTHOR]

Abstract This paper investigates the passive pulse period for the passive neuron in discrete PCNN. We first define a dynamic comparative ratio instead of the logical comparison to describe the linear difference between neural inner state and dynamic threshold. Then a nearly accurate formula about the passive pulse period is given by using the max lower limit of dynamic comparative ratios, and the rationality of which is proved based on the error analysis between estimated and real passive pulse periods. Moreover, we deduce a stable pulse period from estimated pulse period such that the neuron could nonperiodically and periodically pulse in two different time phases, successively. Further, the initial phase, from which the passive neuron can start to pulse periodically, is estimated. Some examples are performed, and the results reach the consensus with theoretical analyses. [ABSTRACT FROM AUTHOR]

Abstract This paper deals with the problem of delay-dependent robust dissipativity analysis for uncertain neural networks with additive time varying delays by using a more general activation function approach. Different from previous literature, some sufficient information on neuron activation function and additive time-varying delays have been considered. By constructing suitable Lyapunov–Krasovskii functionals (LKFs) with some new integral terms, and estimating their derivative by using newly developed single integral inequality that includes Jensen's inequality and Wirtinger-based integral inequality as a special case. A new delay-dependent less conservative global asymptotic stability and dissipative criteria have been established in the form of linear matrix inequalities (LMIs) technique. The effectiveness and advantages of the proposed results are verified by available standard numerical packages. [ABSTRACT FROM AUTHOR]

Abstract In this paper, synchronization results for semi-Markovian jump neural networks with time-varying delays are investigated based on the event-triggered control scheme. With the construction of suitable Lyapunov–Krasovskii functional (LKF), novel synchronization criteria for delayed semi-Markovian jump neural networks are established in the form of linear matrix inequalities (LMIs). Rather, a general free-weighting matrix approach, which is proven to produce less conservative results than the existing methods, is employed to estimate the single integral term. The desired synchronization is achieved by solving the obtained set of LMIs. Eventually, numerical examples are proposed to show the validity of the proposed approach. [ABSTRACT FROM AUTHOR]

Abstract This article deals with the robust passivity analysis problem for uncertain neural networks with both leakage delay and additive time-varying delays by using a more general activation function technique. The information of activation function which is ignored in the existing results is taken into account in this paper. Based on Lyapunov stability theory, a proper Lyapunov–Krasovskii functional (LKF) with some new terms is constructed. The less conservative delay-dependent stability criteria have been obtained by applying a newly developed integral inequality that includes Jensen's inequality and a Wirtinger-based integral inequality as a special case. Some sufficient conditions are achieved to guarantee the stability and passivity of the addressed system model. All the proposed results are formulated as linear matrix inequalities (LMIs). Finally, three numerical cases are simulated to show the effectiveness and benefits of our proposed method. [ABSTRACT FROM AUTHOR]

The paper discusses global Lagrange stability for a class of Cohen–Grossberg BAM neural networks of neutral-type with multiple time-varying and finite distributed delays by constructing appropriate Lyapunov-like functions. To this end, we first establish a new differential integral inequality for non-autonomous Cohen–Grossberg BAM neural networks with finite distributed delays. By using the new inequality and employing some other inequality techniques, we analyze two different types of activation functions which include both bounded and unbounded activation functions. Some easily verifiable criteria are obtained for global exponential attractive sets in which all trajectories converge. These results can also be applied to analyze monostable as well as multistable and more extensive neural networks due to making no assumptions on the number of equilibria. Meanwhile, the results obtained in this paper are more general and challenging than that of the existing references. Finally, some examples with numerical simulations are given and analyzed to verify our results. [ABSTRACT FROM AUTHOR]

Abstract: Without assuming the symmetry or irreducibility of the coupling configuration matrices, this paper investigates the pinning synchronization in an array of linearly coupled delayed neural networks with both constant and delayed couplings. Under some low-dimensional conditions, it is shown that the network can be pinned to a homogenous state by applying adaptive feedback control actions to a small fraction of network nodes. In particular, the paper addresses how to determine a set of pinned nodes for a linearly coupled delayed neural network model with fixed coupling matrices and strengths. Numerical simulations are given to verify the theoretical analysis. [Copyright &y& Elsevier]

Abstract: In this paper, the existence and local exponential stability of the almost periodic solutions for recurrent neural networks with mixed delays have been investigated. By applying Dini derivative and introducing many real parameters, and estimating the upper bound of solutions of the system, a series of new and useful criteria on the existence and local exponential stability of almost periodic for general delayed neural networks without global Lipschitz activation functions have been derived. Those results obtained in this paper extend and generalize the corresponding results existing in the previous literature. Two examples and numerical simulations are given to illustrate our theory. [Copyright &y& Elsevier]

Abstract: This paper studies stationary oscillation for a time-varying recurrent cellular neural network with time delays and impulses. In a recent paper, the authors claim that they obtain a criterion of existence, uniqueness, and global exponential stability of periodic solution (i.e. stationary oscillation) for a recurrent cellular neural network with time delays and impulses. We point out that the main result of their paper is incorrect, and present a sufficient condition of stationary oscillation for a time-varying recurrent cellular neural networks with time delays and impulses. An numerical example is given to illustrate the effectiveness of the obtained result. [Copyright &y& Elsevier]

Abstract: In this paper Hopfield neural networks with continuously distributed delays are considered. Without assuming the global Lipschitz conditions of activation functions, sufficient conditions for the existence and exponential stability of the almost periodic solutions are established by using the fixed point theorem and differential inequality techniques. The results of this paper are new and they complement previously known results. [Copyright &y& Elsevier]

Mobile robots need to navigate in their environment in order to perform useful tasks. Doors appear in almost every office-like indoor environment and they have to be crossed often during the navigation process. We present in this paper a new approach that uses visual information to anticipate that a door has to be crossed. Combining then visual information with ultrasonic sensors, the robot approaches the door until an adequate distance is reached. Door traversing is then performed using sonar sensors. This paper describes the control architecture and the behaviors that have been implemented to obtain the door traversing behavior. Results and performance issues are explained. The experiments have been carried out with a B21 mobile robot. [Copyright &y& Elsevier]

This paper deals with dynamic behaviors of Hopfield type neural network model of n ( ≥ 3 ) identical neurons with two time-delayed connections coupled in a star configuration. Delay dependent as well as independent local stability conditions about trivial equilibrium is found. Considering synaptic weight and time delay as parameters Hopf-bifurcation, steady-state bifurcation and equivariant steady state bifurcation criteria are given. The criterion for the global stability of the system is presented by constructing a suitable Lyapunov functional. Also conditions for absolute synchronization about the trivial equilibrium are also shown. Using normal form method and the center manifold theory the direction of the Hopf-bifurcation, stability and the properties of Hopf-bifurcating periodic solutions are determined. Numerical simulations are presented to verify the analytical results. The effect of synaptic weight and delay on different types of oscillations, e.g. in-phase, phase-locking, standing wave and oscillation death, has been shown numerically. [ABSTRACT FROM AUTHOR]

This paper proposes a mathematical model of multidirectional associative memory (MAM) neural network with almost periodic coefficients and continuously distributed delays. By using exponential dichotomy and spectral radius, a sufficient condition for the existence and the global exponential stability of an almost periodic solution is obtained. Two examples are given to illustrate the effectiveness of our results. [ABSTRACT FROM AUTHOR]

Abstract: This paper introduces the new ATISBAT model for generating accelerated-time simulations of baggage handling in an airport terminal. The philosophy of this model combines ideas from cellular automaton and neural network theories. The basics of this model was presented at ACA’2012 (in the Nonstandard Applications of Computer Algebra Session). The algorithms involved in the ATISBAT model have been implemented using a Computer Algebra System since some symbolic computations are needed. On the other hand, a Java application that shows the simulation graphically has also been developed. One of the most important characteristics of ATISBAT is its flexibility in two different ways. It allows for the simulation of a wide variety of baggage handling systems. It also allows users to choose from among a great list of random distributions (or even create an “ad hoc” one) to simulate the behavior of the random events within the system. The accelerated-time simulations obtained with ATISBAT, can be used both as an aid during the terminal's design process (in order to optimize the baggage handling network's topology) and also to analyze the impact of any possible proposed changes in an already built baggage handling system. The graphic interface developed produces important visual information about the simulation. This approach is very useful, since the effects of making any change to the system can be immediately visualized. [Copyright &y& Elsevier]

Abstract: In this paper we introduce a new model, called the GRAM model, that provides an accelerated-time simulation of traffic on a motorway. The model provides an accelerated-time simulation using ideas and techniques from both cellular automata and neural network models. It has been implemented in a package using a Computer Algebra System (CAS) for the development of algorithms and a Java environment to display the simulated results graphically. The graphic interface shows the actual state of the cars in the system and provides visual information about the current situation of the traffic on the motorway. Therefore, this graphical approach can be used to help in the process of designing the traffic control system since the effects of making any change can be seen immediately. As an example of using GRAM, some simulations have been generated in a stretch of motorway where two different designs (with and without an accelerating lane) have been considered. As a result, it is possible to analyse the results of including an accelerating lane on the number of traffic jams on the motorway. [Copyright &y& Elsevier]

Abstract: A new neural network is proposed in this paper for solving quadratic programming problems with equality and inequality constraints. Comparing with the existing neural networks for solving such problems, the proposed neural network has fewer neurons and an one-layer architecture. The proposed neural network is proven to be global convergence. Furthermore, illustrative examples are given to show the effectiveness of the proposed neural network. [Copyright &y& Elsevier]

Abstract: In this paper, we present multi-almost periodicity of a general class of discrete-time neural networks derived from a well-known semi-discretization technique, that is, coexistence and exponential stability of 2 N almost periodic sequence solutions of discrete-time neural networks subjected to external almost periodic stimuli. By using monotonicity and boundedness of activation functions, we construct 2 N close regions to attain the existence of almost periodic sequence solutions. Meanwhile, some new and simple criteria are derived for the networks to converge exponentially toward 2 N almost periodic sequence solutions. As special cases, our results can extend to discrete-time analogues of periodic or autonomous neural networks and hence complement or improve corresponding existing ones. Finally, computer numerical simulations are performed to illustrate multi-almost periodicity of semi-discretizations of neural networks. [Copyright &y& Elsevier]

Abstract: The paper presents a method for FPGA implementation of Self-Organizing Map (SOM) artificial neural networks with on-chip learning algorithm. The method aims to build up a specific neural network using generic blocks designed in the MathWorks Simulink environment. The main characteristics of this original solution are: on-chip learning algorithm implementation, high reconfiguration capability and operation under real time constraints. An extended analysis has been carried out on the hardware resources used to implement the whole SOM network, as well as each individual component block. [Copyright &y& Elsevier]

Abstract: Two new methods to identify the mechanical parameters in induction motor based field oriented drives are presented in this paper. The identified parameters are: the moment of inertia and the viscous damping coefficient. The proposed methods are based on the adaptive linear neuron (ADALINE) networks. The two parameters are derived and optimized during the online training process. During the identification phase, the motor torque is controlled by the well-known field oriented control strategy. This torque is subjected to variations in order to obtain mechanical speed transients. The two proposed methods are simple to implement compared to the previous techniques. They require only the stator current and mechanical speed measurements. Finally, the effectiveness of the two methods and the accuracy of the derived parameters are proven experimentally by two direct starting tests. The originality of this work is the building of a model representation that it is suitable for implementation with ADALINE networks. This leads to a simple implementation and ease of mechanical parameters identification. [Copyright &y& Elsevier]

Abstract: The development of the Financial Crisis throughout 2008 and 2009 has made many investors and fund managers question whether growth-based investment approaches have had their day. Value-based approaches built on fundamental analysis have resurfaced again. Typically, these value-based models use fundamental variables to decide between investment opportunities. In a previous work, Vanstone et al. studied a set of filters published by Aby et al. during the dot-com crash of 2000 and subsequent aftermath, and tested and benchmarked these filters in the Australian market. The Aby filters rely on 4 different fundamental variables, and use rules with specific cut-off values to determine when to enter and exit trades. These cut-off values were found to be too restrictive for the Australian markets. This paper uses a neural network methodology by Vanstone and Finnie to develop a stockmarket trading system based on these same 4 fundamental variables, and demonstrates the important role neural networks have to play within complex and noisy environments, such as that provided by the stockmarket. [Copyright &y& Elsevier]

Abstract: The paper studies the dynamical model of motor cognition of neural networks through the theory of stochastic phase resetting dynamics, presents the interaction, phase coding, and the evolution of the time-varying averaged number density in terms of populations of perceptive neurons, inter-neurons, and motor neurons subject to coupling, and probes into the dynamical reaction of neural networks under the condition of spontaneous movement and stimulation, respectively. With numerical simulations, we prove (1) Walter J. Freeman’s conjecture that the response of cortex dynamics cannot code external stimulation information; (2) the possession of rhythm coding in the neural coding of serial neural networks; (3) the importance of neural inhibition in the regulation of the central nervous system. [Copyright &y& Elsevier]

Abstract: This paper investigates the asymptotic stability of a class of impulsive high-order neural networks, which can be considered as an expansion of Hopfield neural networks. By employing Lyapunov functions and linear matrix inequality (LMI) technique, sufficient conditions that guarantee the global asymptotic stability of the equilibrium point are derived. The proposed criteria are easily verified and possess many adjustable parameters, which provide flexibility for the analysis of the neural networks. Finally, two examples are given to show the effectiveness of the proposed results. [Copyright &y& Elsevier]

Abstract: In this paper, a generalized recurrent neural network is proposed for solving ϵ-insensitive support vector regression (ϵ-ISVR). The ϵ-ISVR is first formulated as a convex non-smooth programming problem, and then a generalize recurrent neural network with lower model complexity is designed for training the support vector machine. Furthermore, simulation results are given to demonstrate the effectiveness and performance of the proposed neural network. [Copyright &y& Elsevier]

Abstract: In this paper, we show that noise can transform a hybrid neural networks, whose solution may grow exponentially, into a new stochastic one, whose solution grows at most polynomially. In other words, we reveal that noise can suppress the exponential growth in hybrid Hopfield neural networks. [Copyright &y& Elsevier]

Abstract: In this paper, a class of non-autonomous reaction-diffusion neural networks with time-varying delays is investigated. By establishing a new differential inequality and employing the properties of spectral radius of nonnegative matrix and diffusion operator, the global attracting and positive invariant sets and exponential stability of non-autonomous reaction-diffusion neural networks with time-varying delays are obtained. Our results do not require the conditions of boundedness of the coefficient of neural networks. One example is given to illustrate the effectiveness of our conclusion. [Copyright &y& Elsevier]

Abstract: In this paper, we establish a method to study the mean square exponential stability of the zero solution of impulsive stochastic reaction-diffusion Cohen–Grossberg neural networks with delays. By using the properties of M-cone and inequality technique, we obtain some sufficient conditions ensuring mean square exponential stability of the zero solution of impulsive stochastic reaction-diffusion Cohen–Grossberg neural networks with delays. The sufficient conditions are easily checked in practice by simple algebra methods and have a wider adaptive range. Two examples are also discussed to illustrate the efficiency of the obtained results. [Copyright &y& Elsevier]

Abstract: In this paper, lag synchronization for a class of delayed fuzzy cellular networks is investigated. By utilizing inequality technique, Lyapunov functional theory and the analysis method, some new and useful criteria of lag synchronization for the addressed networks are derived in terms of p-norm under a periodically intermittent controller. Finally, an example with simulation is given to show the effectiveness of the obtained results. [Copyright &y& Elsevier]

Abstract: This paper presents an integrated artificial neural network-computer simulation (ANNSim) for optimization of G/G/K queue systems. The ANNSim is a computer program capable of improving its performance by referring to production constraints, system''s limitations and desired targets. It is a goal oriented, flexible and integrated approach and produces the optimum solution by utilizing Multi Layer Perceptron (MLP) neural networks. The properties and modules of the prescribed intelligent ANNSim are: (1) parametric modeling, (2) flexibility module, (3) integrated modeling, (4) knowledge-base module, (5) integrated database and (6) learning module. The integrated ANNSim is applied to 30 distinct tandem G/G/K queue systems. Furthermore, its superiority over conventional simulation approach is shown in two dimensions which are average run time and maximum number of required iterations (scenarios). [Copyright &y& Elsevier]

Abstract: This paper presents an overview of an analysis method based on self-organizing maps (SOM) which was applied to an activated sludge treatment process in a pulp mill. The aim of the study was to determine whether the neural network modeling method could be a useful and time-saving way to analyze this kind of process data. The following analysis procedure was used. At first, the process data was modeled using the SOM algorithm. Next, the reference vectors of the map were classified by K-means algorithm into clusters, which represented different states of the process. At the final stage, the reference vectors of the map and the centre vectors of the clusters were used for subtraction analysis to indicate differences of the process states. The results show that the method presented here can be an efficient way to analyze the data of an activated sludge treatment process. [Copyright &y& Elsevier]

Abstract: In this paper, we investigate nonlinear adaptive–impulsive synchronization of chaotic systems. The core of adaptive–impulsive synchronization of chaotic systems is to design suitable nonlinear adaptive–impulsive controllers to attain certain aims. Based on invariant principle of impulsive dynamical systems, nonlinear adaptive–impulsive controllers are derived to make the state of chaotic systems synchronized. Accordingly, concrete nonlinear adaptive–impulsive control scheme is designed for quantum cellular neural network (Quantum-CNN). Moreover, numerical simulations are presented to illustrate the effectiveness of the proposed nonlinear controllers, which is more effective than that in the previous literature. [Copyright &y& Elsevier]

Abstract: Many fuzzy time series approaches have been proposed in recent years. These methods include three main phases such as fuzzification, defining fuzzy relationships and, defuzzification. Aladag et al. improved the forecasting accuracy by utilizing feed forward neural networks to determine fuzzy relationships in high order fuzzy time series. Another study for increasing forecasting accuracy was made by Cheng et al. . In their study, they employ adaptive expectation model to adopt forecasts obtained from first order fuzzy time series forecasting model. In this study, we propose a novel high order fuzzy time series method in order to obtain more accurate forecasts. In the proposed method, fuzzy relationships are defined by feed forward neural networks and adaptive expectation model is used for adjusting forecasted values. Unlike the papers of Cheng et al. and Liu et al. , forecast adjusting is done by using constraint optimization for weighted parameter. The proposed method is applied to the enrollments of the University of Alabama and the obtained forecasting results compared to those obtained from other approaches are available in the literature. As a result of comparison, it is clearly seen that the proposed method significantly increases the forecasting accuracy. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, the exponential stability analysis for the bidirectional associative memory neural network model with both time-varying delays and general activation functions is considered. Neither the boundedness and the monotony on these activation functions nor the differentiability on the time-varying delays are assumed. By employing Lyapunov functional and the linear matrix inequality (LMI) approach, several new sufficient conditions in LMI form are obtained to ensure the existence, uniqueness and global exponential stability of equilibrium point for the neural networks. Moreover, the exponential convergence rate index is estimated, which depends on the system parameters. The proposed stability results are less conservative than some recently known ones in the literature, which is demonstrated via an example with simulation. [ABSTRACT FROM AUTHOR]